QUANTIFYING THE EFFECTS OF SAMPLING METHOD AND EFFORT ON ESTIMATES OF SPECIES RICHNESS AND RELATIVE ABUNDANCE OF NEOGENE BENTHIC MARINE MOLLUSKS (CIBAO VALLEY, DOMINICAN REPUBLIC)

This study is the first phase of a large-scale project that is investigating how best to sample, quantify and compare patterns of community stability and change through time in the extensively studied and sampled Neogene stratigraphic sections of the Cibao Valley, northern Dominican Republic. Our study focuses on how sampling methods (bulk vs. float) and effort (number of bulk samples) influence quantitative estimates of species richness and relative abundance, which are important variables for the delineation of ecological and paleoecological communities. Our research focused on locality NMB 15805 (Early Pliocene, NN12, Gurabo Formation). We collected 6 replicate bulk samples (2.6 Kg, 2.8 liters each) laterally along a shell bed at this locality. For float sampling, we collected all visible macrofossils at the outcrop surface (0-1.53 m vertical, 80 m horizontal). We hypothesized that bulk and float sampling would produce equivalent estimates of species richness and relative abundance in the same mollusk-rich shell bed at the same locality. Rarefaction analyses (with bootstrapped error estimates) of species abundance data indicate that species richness is not statistically different between bulk and float samples (at n=110). In rarefaction analyses of bulk samples (at n=310), most bulks had statistically indistinguishable species richness values. However, some significant differences in pair-wise comparisons of species richness were found between bulks. Quantitative comparisons of rank order abundance, however, indicate that bulk and float samples produced statistically different results. In contrast, similar rank order abundances occur among bulk samples. These results have significant implications for community reconstruction and delineation in future studies of the Gurabo section. Past collections by the Naturhistorishes Museum Basel appear to be very poorly suited for identifying paleocommunities using rank order abundances because a significant portion of the material was collected as float.